1. What is the WHT anyway, and why are there so many ways to compute it? Jeremy Johnson 1, 2, 6, 24, 112, 568, 3032, 16768,…

2. Walsh-Hadamard Transform y = WHT N x, N = 2 n

3. WHT Algorithms Factor WHT N into a product of sparse structured matrices Compute: y = (M 1 M 2 … M t )x y t = M t x … y 2 = M 2 y 3 y = M 1 y 2

4. Factoring the WHT Matrix AC  D  C  D  A  A   C) = ( A   C I m  n  mn WHT 2  WHT 2  WHT 2      WHT 2 

5. Recursive and Iterative Factorization WHT 8  WHT 2      WHT 4   WHT 2      WHT 2     I 2  WHT    WHT 2      WHT 2     I 2  I 2  WHT    WHT 2      WHT 2     I 2  I 2  WHT    WHT 2      WHT 2     I 2  I 2  WHT    WHT 2      WHT 2     I 4  WHT  

6. WHT 8 =  WHT 2      WHT 2     I 4  WHT  

7. WHT Algorithms Recursive Iterative General

8. WHT Implementation Definition/formula –N=N 1 * N 2  N t N i =2 n i –x=WHT N *x x =(x(b),x(b+s),…x(b+(M-1)s)) Implementation(nested loop) R=N; S=1; for i=t,…,1 R=R/N i for j=0,…,R-1 for k=0,…,S-1 S=S* N i; M b,s t i1 )     nn WHT 2 2 2 II( n 1 + ··· +n i- 1 2 n i+1 + ··· +n t i

9. Partition Trees 4 13 12 11 4 1111 Right Recursive Iterative 9 342 121 11 4 13 12 11 Left Recursive 4 22 1111 Balanced

10. Ordered Partitions There is a 1-1 mapping from ordered partitions of n onto (n-1)-bit binary numbers.  There are 2 n-1 ordered partitions of n. 1|1 1|1 1 1 1|1 1  1+2+4+2 = 9 162 = 1 0 1 0 0 0 1 0

11. Enumerating Partition Trees 33 12 0001 3 12 11 3 21 10 3 12 11 3 11 11 1

12. Counting Partition Trees

13. WHT Package Püschel & Johnson (ICASSP ’00) Allows easy implementation of any of the possible WHT algorithms Partition tree representation W(n)=small[n] | split[W(n 1 ),…W(n t )] Tools –Measure runtime of any algorithm –Measure hardware events (coupled with PCL) –Search for good implementation Dynamic programming Evolutionary algorithm

14. Histogram (n = 16, 10,000 samples) Wide range in performance despite equal number of arithmetic operations (n2 n flops) Pentium III consumes more run time (more pipeline stages) Ultra SPARC II spans a larger range

15. Operation Count Theorem. Let W N be a WHT algorithm of size N. Then the number of floating point operations (flops) used by W N is Nlg(N). Proof. By induction.

16. Instruction Count Model A(n) = number of calls to WHT procedure  = number of instructions outside loops A l (n) = Number of calls to base case of size l  l = number of instructions in base case of size l L i = number of iterations of outer (i=1), middle (i=2), and outer (i=3) loop  i = number of instructions in outer (i=1), middle (i=2), and outer (i=3) loop body

17. Small[1].file"s_1.c".version"01.01" gcc2_compiled.:.text.align 4.globl apply_small1.type apply_small1,@function apply_small1: movl 8(%esp),%edx //load stride S to EDX movl 12(%esp),%eax //load x array's base address to EAX fldl (%eax) // st(0)=R7=x[0] fldl (%eax,%edx,8) //st(0)=R6=x[S] fld %st(1) //st(0)=R5=x[0] fadd %st(1),%st // R5=x[0]+x[S] fxch %st(2) //st(0)=R5=x[0],s(2)=R7=x[0]+x[S] fsubp %st,%st(1) //st(0)=R6=x[S]-x[0] ????? fxch %st(1) //st(0)=R6=x[0]+x[S],st(1)=R7=x[S]-x[0] fstpl (%eax) //store x[0]=x[0]+x[S] fstpl (%eax,%edx,8) //store x[0]=x[0]-x[S] ret

18. Recurrences

20. Histogram using Instruction Model (P3)  l = 12,  l = 34, and  l = 106  = 27  1 = 18,  2 = 18, and  1 = 20

21. Algorithm Comparison

22. Dynamic Programming min n 1 +… + n t = n Cost ( TT … n1n1 ntnt n ), where T n is the optimial tree of size n. This depends on the assumption that Cost only depends on the size of a tree and not where it is located. (true for IC, but false for runtime). For IC, the optimal tree is iterative with appropriate leaves. For runtime, DP is a good heuristic (used with binary trees).